http://www.mrc-cbu.cam.ac.uk/personal/daniel.bor/pvs2002.pdf
Neurocase (2002) Vol. 8, pp. 394–403 (c) Oxford University Press 2002
Detecting Residual Cognitive Function in Persistent
Vegetative State
Adrian M. Owen, David K. Menon1, Ingrid S. Johnsrude, Daniel Bor,
Sophie K. Scott2, Tom Manly,
Emma J. Williams1, C. Mummery3 and John D. Pickard1
MRC Cognition and Brain Sciences Unit, 15 Chaucer Road, Cambridge CB2
2EF, 1Wolfson Brain Imaging Centre,
Addenbrooke's Hospital, University of Cambridge, Hills Road, Cambridge
CB2 2QQ, 2Department of Psychology, University
College London, Gower Street, London WC1E 6BT and 3Wellcome Department
of Cognitive Neurology, 12 Queen Square,
London, WC1N 3BG, UK
Abstract
Despite converging agreement about the definition of persistent
vegetative state, recent reports have raised concerns
about the accuracy of diagnosis in some patients, and the extent to
which, in a selection of cases, residual cognitive
functions may remain undetected. Objective assessment of residual
cognitive function can be extremely difficult as
motor responses may be minimal, inconsistent, and difficult to
document in many patients, or may be undetectable in
others because no cognitive output is possible. Here we describe
strategies for using H2
15O positron emission
tomography activation studies to study covert cognitive processing in
patients with a clinical diagnosis of persistent
vegetative state. Three cases are described in detail. Of these, two
exhibited clear and predicted regional cerebral blood
flow responses during well-documented activation paradigms (face
recognition and speech perception) which have
been shown to produce specific, robust and reproducible activation
patterns in normal volunteers. Some months after
scanning, both patients made a significant recovery. In a third case,
blood flow data were acquired during a speech
perception task, although methodological difficulties precluded any
systematic interpretation of the results. In spite of
the multiple logistic and procedural problems involved, these results
have major clinical and scientific implications and
provide a strong basis for the systematic study of possible residual
cognitive function in patients diagnosed as being
in a persistent vegetative state.
Introduction
Persistent vegetative state (PVS) was formally defined by
Jennett and Plum (1972) and described as a state of 'wakefulness
without awareness'. Aetiology is variable, although
the condition may arise as a result of a road traffic accident,
ischaemic attack, anoxia, encephalitis or viral infection. A
diagnosis of PVS is not normally considered until between
1 and 3 months post-ictus at which point there must be no
evidence of sustained, reproducible, purposeful or voluntary
behavioural response to visual, auditory, tactile or noxious
stimuli. There must also be no evidence of language comprehension
or expression, although there is generally sufficiently
preserved hypothalamic and brain stem autonomic functions
to permit survival with medical care. Although PVS often
follows coma, it is characterized by an irregular but cyclic
state of circadian sleeping and waking. In contrast, patients
in coma present with eyes closed and lack any consistent
sleep–wake cycles.
The majority of imaging studies in patients with PVS have
used either fluorodeoxyglucose (FDG) positron emission
Correspondence to: Adrian M. Owen, MRC Cognition and Brain Sciences
Unit, 15 Chaucer Road, Cambridge CB2 2EF, UK. Tel: 44 (0)1223 355294
(ext. 511); Fax: 44 (0)1223 359062; e-mail: [EMAIL PROTECTED]
tomography (PET) or single photon emission computed
tomography (SPECT) and have reported widespread reductions
of up to 50% in (resting) cerebral blood flow and
glucose metabolism (Levy et al., 1987; DeVolder et al., 1994;
Tommasino et al., 1995). In some cases, however, isolated
'islands' of metabolism have been identified in circumscribed
regions of cortex, which may suggest residual cognitive
processing in a subset of patients (Schiff and Plum, 1999).
While metabolic studies are useful in this regard, they can
only identify functionality at the most general level; that is,
mapping cortical and subcortical regions that are potentially
recruitable, rather than relating neural activity within such
regions to specific cognitive processes (Momose et al., 1989;
Turkstra, 1995).
To some extent, electrophysiological studies do not suffer
this same limitation and have been applied successfully to
the problem of PVS (Turkstra, 1995), although with the
exception of magnetoencephalography (MEG), these
approaches lack sufficient spatial resolution to assess fully
Residual cognitive function in PVS 395
cognitive function in these patients. Unfortunately, MEG is
still not widely available and, in any case, comparative data
from healthy volunteers remain sparse.
So-called 'activation studies' using H2
15O PET or functional
magnetic resonance imaging (fMRI) together with
established sensory paradigms may provide a viable method
for assessing cognitive processing in patients with PVS. In
short, given the unique problem of assessing patients with
PVS, functional neuroimaging has the potential to demonstrate
distinct and specific physiological responses [changes
in regional cerebral blood flow (rCBF) or changes in regional
cerebral haemodynamics] to controlled external stimulation
in the absence of any overt response on the part of the
patient. The technique does, however, pose a number of
methodological, ethical and procedural problems. For
example, motor responses are often minimal, inconsistent or
absent in patients with PVS and by definition cannot be
elicited directly (e.g. willfully) by external stimulation. In
addition, even assuming that some level of residual cognitive
processing does exist, there is no reliable mechanism for
ensuring that the presented stimuli are actually perceived by
the patient. Many PVS patients suffer serious damage to
auditory and/or visual input systems, which may impede
performance of any 'higher' cognitive functions (e.g. voice
discrimination) which place demands on these 'lower' sensory
systems (e.g. hearing). Like patients with any form of serious
brain damage, PVS may also be accompanied by a significant
reduction in attention span (assuming some level of cognitive
processing remains), which may further complicate the assessment
of higher cognitive functions. Spontaneous movements
during the scan itself may also compromise the interpretation
of functional neuroimaging data, particularly scans acquired
using fMRI. Where PET methodology is employed, issues
of radiation burden must also be considered and may preclude
longitudinal or follow-up studies in many patients. Finally,
data processing of functional neuroimaging data may also
present challenging problems in patients with PVS. For
example, the presence of gross hydrocephalus or focal pathology
may complicate co-registration of functional data (e.g.
acquired with PET or fMRI) to anatomical data (e.g. acquired
using structural MRI), and the normalization of images to a
healthy reference brain. Under these circumstances statistical
assessment of activation patterns is complex and interpretation
of activation foci in terms of standard stereotaxic
coordinates may be impossible.
In this study we used H2
15O PET to study covert cognitive
processing in three patients with a probable clinical diagnosis
of PVS. In the first case, a decision was made to use visual
(face) stimuli on the basis of a preliminary PET study that
demonstrated activation in primary visual cortex (V1) in
response to moving coloured visual stimuli on a computer
screen. The decision was reinforced further by non-reproducible
reports of visual pursuit in response to faces of
family and friends. In the second patient, an auditory (speech)
task was employed on the basis of anecdotal reports of
occasional movement following verbal commands and intact
brain auditory evoked responses on one side. In a third patient,
speech recognition was also employed as the paradigm
for investigation, but procedural complications during the
scanning session precluded the acquisition of any useful data.
Materials and methods
Case histories
Case 1 (KB) was a 26-year-old female who presented with
an acute febrile illness which culminated in a depression of
her conscious state. MRI revealed widespread hyperintensity
in the brain stem, both thalami, and anterior and medial
temporal lobes. At first assessment, 4 months after admission,
her eyes opened spontaneously and she exhibited sleep–wake
cycles. Anecdotal evidence suggested that she occasionally
followed family members with her eyes, but despite repeated
examination, the patient showed no consistent spontaneous
or elicited motor responses or eye movements to suggest that
she could communicate. The pons and mid-brain components
of brain stem auditory evoked responses were abnormal, but
a delayed auditory oddball P300 could be detected. A
diagnosis of probable PVS was made. Clinical findings and
examination of cerebrospinal fluid were consistent with acute
disseminated encephalomyelitis. MRI showed hyperintensity
in the brain stem and small foci of hyperintensity in both
thalami and in the medial right temporal lobe on T2-weighted
images. When assessed the patient had a tracheotomy, was
fed through a gastrotomy and was doubly incontinent.
The second patient (DE), a 30-year-old female bank
manager, suffered severe head injuries during a road traffic
accident involving a head-on collision with another vehicle.
During a significant period trapped in the car, she probably
suffered a period of hypoxia with hypotension. The Glasgow
Coma Scale at the scene of the accident indicated a score of
4/15 with no improvement post-resuscitation. A long stay in
intensive care was accompanied by episodes of pupillary
unreactivity on more than one occasion. Fourteen weeks
post-ictus, the pupils remained dilated and unreactive. The
patient was weaned off a ventilator but still required a
tracheotomy. Brain stem auditory evoked responses were
intact on the right but abnormal on the left. Computed
tomography (CT) findings at admission revealed a left frontal
subcortical haemorrhagic contusion and a smaller frontoparietal
contusion. A small haemorrhage in the inter-peduncular
fossa with low-density areas in the adjacent mid-brain–pons
region were also observed. A repeat scan on the same day
showed fresh midline haemorrhage in the anterior midbrain
extending to posteromedial right thalamus. In addition,
punctuate areas of high density in the right cerebellum and
both cerebral hemispheres suggested diffuse axonal injury.
The cerebral ventricles were noted to be normal in size. Over
several weeks she developed a withdrawal to pain but showed
no consistent evidence of volitional activity. In spite of this,
the family felt that, when not tired, the patient occasionally
396 A. M. Owen et al.
showed responses to commands. She was doubly incontinent,
fed through a gastrotomy and required 24-h nursing care.
The third patient (JB) was a 28-year-old female with
probable mixed connective tissue disease treated with
immunosuppression. She suffered cardiorespiratory arrest
during an episode of acute deterioration and was left in a
PVS. Four months post-ictus the patient exhibited no signs
of neurological awareness, was fed via a nasogastric tube and
was incontinent with no voluntary movements or responses to
external stimuli.
Informed written consent for participation was obtained
for each patient from the next-of-kin after the nature of the
study and possible consequences had been fully explained.
The study was approved by the Cambridgeshire Local
Research and Ethics Committee.
Control subjects
A healthy, 30-year-old, male (TM), control volunteer with
English as his first language and no history of neurological
or neuropsychiatric disturbance or substance abuse, was
selected for comparison with patient KB. Local ethical
considerations precluded the testing of a young female (i.e.
gender-matched) volunteer. The control subject underwent
precisely the same scanning conditions as patient KB (see
below), and the data were analysed using an identical
procedure (see Scanning methods). Informed, written consent
for participation in the study was obtained after its nature
and possible consequences had been explained to him.
It was unfortunately not possible to acquire single-subject
matched control data for the two remaining patients within
the context of this preliminary investigation. However, the
paradigm employed in these two patients is very well
established and has been used in previous studies in healthy
control volunteers. Therefore, the data from these two patients
were compared with that published previously on six normal
healthy male control subjects (age range 26–58 years) using
an identical paradigm (Mummery et al., 1999). Full details
of that study will not be replicated here, although all subjects
had English as their first language and none had any significant
history of neurological or neuropsychiatric disturbance or
substance abuse.
Image acquisition and data analysis
In patients KB, DE and JB and in the normal control volunteer
TM, PET scans were obtained with the General Electrics
Advance system, which produces 35 image slices at an
intrinsic resolution of approximately 4.0 5.0 4.5 mm.
Patient KB underwent two separate PET investigations (see
below) comprising 12 and 15 scans, respectively, while DE,
JB and TM were each scanned within a single PET session
comprising 12 scans. Using the bolus H2
15O methodology,
rCBF was measured during the 12 separate scans (or 15 in
the case of the second session for patient KB, see below).
For each scan, the subjects received a 20-s intravenous bolus
of H2
15O through a forearm cannula at a concentration of
300 Mbq/ml and a flow rate of 10 ml/min. With this method,
each scan provides an image of rCBF integrated over a
period of 90 s from when the tracer first enters the cerebral
circulation. The 12 PET scans were realigned using the first
scan as a reference, normalized for global CBF value and
averaged within each patient for each activation state (experimental
task and control task). The images were then smoothed
using an isotropic Gaussian kernel at 16 mm. Finally, a
simple ANCOVA (analysis of covariance) model was fitted
to the data at each voxel, as implemented by the method of
statistical parametric mapping (SPM 96, provided by the
Wellcome Department of Cognitive Neurology, London, UK),
with a condition effect for each of the conditions, using
global CBF as a confounding covariate. For each patient, a
three-dimensional MRI volume (256 256 128 pixels,
3 mm thick) was acquired and resliced so as to be coregistered
with the PET data. The co-registered PET data
were transformed into standard stereotaxic space using conventional
algorithms within the SPM 96 analysis environment.
The significance of a given rCBF difference was assessed by
application of an intensity threshold to the SPM images
(Worsley et al., 1992, 1996). This threshold, based on
three-dimensional Gaussian random field theory, predicts the
likelihood of obtaining a false positive in an extended threedimensional
field.
Stimuli and testing conditions
Across the patients and control subject, three separate paradigms
were employed and are described below.
Visual stimulation (patient KB only). This task was used
during a pilot study of patient KB, the purpose of which was
to ascertain whether the logistic problems associated with
PET scanning PVS patients could be overcome and, more
importantly, whether any significant changes in rCBF could
be detected. During six scans, a standard Windows 95
screen saver (Microsoft) involving rapidly moving coloured
'windows' was employed. The moving image was presented
on a 20 computer monitor mounted approximately 50 cm
in front of the patient. The patient's eyes remained open for
all six scans. During six further scans, no visual image was
presented and the patient's face was covered with a light
cloth to prevent any visual stimulation. All 12 scans (six
experimental scans and six control scans) were presented in
pseudo-random order. This preliminary PET study was carried
out approximately 1 month before the second study of
familiar face perception, described below.
Familiar face perception (patient KB and control subject
TM). Ten photographs were obtained from the relatives of
patient KB and control subject TM. In neither case was the
participant aware of which photographs had been provided,
although presumably they were all familiar. The photographs
were chosen so as to include, as their main theme, faces of
Residual cognitive function in PVS 397
Fig. 1. Example stimuli from the familiar face perception condition
(left) and the corresponding control task (right). Control pictures
were prepared by distorting
and/or repixellating the experimental stimuli to remove any clear
structure from the images.
friends, family, pets and the participant themselves. The
photographs were digitized at high resolution and were
presented in a large format (approximately 30 cm2) against
a black background on a high-resolution computer monitor.
The monitor was suspended approximately 50 cm above the
subject in the scanner and was therefore placed at a comfortable
viewing distance. A set of 10 control pictures was also
prepared by distorting and/or repixellating the same two sets
of 10 photographs to remove any clear structure from the
images (see Fig. 1). Specifically, the original photographs
were altered such that the overall luminance, colour range,
and solid angle subtended by the pictures and control images
were identical. During the scans the photographs (experimental
condition) or repixellated pictures (control condition)
were presented on the computer screen for 12 s each, starting
30 s before tracer infusion was initiated and continuing until
the end of data acquisition. During the experimental scans
the patient and the control volunteer were told to look at
each of the faces and to 'think about that person'. During
the control scans the subjects were told to 'look at each of
the images'.
Although 12 scans (six experimental and six control) were
planned in patient KB (as in control subject TM), during
three of the scans the patient appeared to close her eyes and
fall asleep. Although the data acquired during these scans
are plotted in Fig. 2 for reference, they were not used in the
comparison of the familiar face perception condition with
the control condition. Three additional scans (making 15 in
total) were conducted in order that a balanced data set was
available for analysis.
Speech perception (patients DE and JB). In the speech
perception tasks the patients were scanned while being
presented binaurally with either spoken words (experimental
condition), or signal-correlated noise stimuli, or no auditory
stimulus at all (rest condition). The presented words were
disyllabic nouns matched for frequency (6–20 000), concreteness
(400–700), and imageability (300–700) on the
Medical Research Council Psycholinguistic Database (Coltheart,
1981), and were pre-recorded on a tape by a speaker,
the rate being controlled by a metronome. The signalcorrelated
noise stimuli were made by selecting a sample of
these spoken nouns with varying segmental durations and
initial manners of articulation, digitizing them and then
multiplying these with noise. The two sets of stimuli were
matched for loudness by adjusting the amplification until
they were subjectively similar. The task instruction was to
'pay attention to the stimuli without responding' or, in the
rest condition, just to rest. In both stimulus conditions,
the words were presented at rates of 30 per minute and
presentations were started 30 s prior to tracer infusion and
continued until the end of data acquisition.
Determination of significance thresholds
Each study was designed to test anatomically specific hypotheses
as both of the tasks used are known to produce welldocumented,
specific, robust and reproducible activation
patterns in normal volunteers. For example, Haxby et al.
(1991) examined rCBF changes while healthy control subjects
performed face matching, dot matching or a sensory motor
control task. Face matching alone activated occipitotemporal
area 37 in the posterior fusiform gyrus, bilaterally. In a
follow-up study (Haxby et al., 1994), the subjects were
required to match faces in one set of scans while in control
scans scrambled patterns of equivalent visual complexity
were shown. The most specific changes in rCBF associated
with face perception (relative to both spatial perception and
perception of scrambled visual stimuli) were observed in
regions of the mid-fusiform and mid-anterior fusiform gyri
(areas 19 and 37).
In spite of this background, single-subject studies using
PET are rare and for this reason we elected to apply a
standard face recognition paradigm, similar to that employed
by Haxby et al. (1991, 1994), to the patient with PVS (KB)
and to the control subject (TM) using an identical procedure
in each case. Accordingly, a significant change in rCBF
during face perception was predicted in the posterior section
of the fusiform gyrus (area 19), particularly in the right
hemisphere. Within this region a directed search was
398 A. M. Owen et al.
conducted and the threshold for reporting a peak as significant
was set at P 0.001, uncorrected for multiple comparisons.
For the rest of the brain, an exploratory search involving all
peaks within the grey matter (volume 600 cm3) was conducted
and the threshold for reporting a peak as significant was set
at P 0.05, corrected for multiple comparisons.
Unfortunately, within the constraints of the clinical investigation,
it was not possible to scan any control subjects
using the speech perception task. However, an identical
investigation has been carried out previously and on the basis
of published findings from that study (Mummery et al., 1999)
significant rCBF changes were predicted in our patients
within the superior temporal cortex of both hemispheres. For
these defined regions the significance threshold was set at
P 0.001, uncorrected. For all other regions, significance
was set at P 0.05, corrected. In all cases only regions
satisfying these criteria are reported.
Results
Visual stimulation (patient KB only)
A comparison of the six control scans (eyes covered, no
stimulation) with the six experimental scans (visual stimulation;
'flying windows' screen saver) yielded one significant
focus of activation in primary visual cortex at coordinates
8, –92, 4 (z 4.59, P 0.031). Intriguingly, a second
activation focus was observed within the right dorsolateral
frontal cortex (45, 34, 21; z 3.91, P 0.348), although
this failed to reach statistical significance according to our
conservative criteria.
Familiar face perception (patient KB and control
subject TM)
In patient KB, subtraction of the six waking scans during
which repixellated control images were presented from the
six waking scans during which familiar faces were presented
yielded just two significant peaks of activation that survived
the statistical threshold described above. Specifically, two
regions within the right fusiform gyrus (Brodmann area 37)
were significantly activated at coordinates 38, –64, 0
(z 3.90, P 0.001) and 44, –66, –20 (z 3.65,
P 0.001) (Fig. 2). The reverse subtraction yielded no
significant foci of activation.
In control subject TM, the equivalent subtraction yielded
a significant activation focus at very similar coordinates (see
Fig. 2) within area 37 of the right fusiform gyrus (34, –76,
–6; z 5.07, P 0.001). In addition, significant rCBF
changes were observed in a similar region within the left
fusiform gyrus (–36, –64, 0; z 5.08, P 0.001), as well
as slightly more anteriorly in area 37 of the same hemisphere
(–28, –44, –10; z 4.76, P 0.016) and in primary visual
cortex (10, –94, 2; z 4.64, P 0.026). The reverse
subtraction yielded no significant foci of activation.
Speech perception (patients DE and JB)
In patient DE, the comparison of signal-correlated noise
stimuli with rest revealed significant foci of activation bilaterally
in the auditory region (–40, –6, –2; z 3.92, P 0.001
and 46, –30, 12; z 3.79, P 0.001), suggesting that
basic auditory processes were at least somewhat functional.
Knowing this, it is possible to compare speech with signalcorrelated
noise stimuli, in the hope of observing activation
that is more specific to speech sounds. The comparison of
speech sounds with signal-correlated noise stimuli revealed
significant rCBF increases on the superior temporal plane
bilaterally (68, –16, –10; z 3.59, P 0.001 and –60,
–12, 2; z 3.01, P 0.001) and posterior to auditory cortex,
in the region of the planum temporale, in the left hemisphere
only (–54, –34, 10; z 3.30, P 0.001) (see Fig. 3). The
reverse subtraction yielded no significant foci of activation.
These findings correspond closely with PET results from a
recent study of six healthy control subjects using identical
stimuli (Mummery et al., 1999). For example, in that study,
significant activation was observed bilaterally on the superior
temporal plane, including auditory areas, extending into the
superior temporal sulcus (see Fig. 3). Although, from this
figure, activation in the Mummery et al. (1999) study appears
to be bilaterally symmetrical in its posterior extent, they
report a unilateral left posterior temporal focus at –54, –38,
8, which is 4 mm from the posterior temporal focus in DE:
this difference is not spatially resolvable with the PET
technique. Zatorre et al. (1992), in another study comparing
speech and noise, have also reported left posterior temporal
activation, although 13 mm more anterior than the focus in
DE and the study of Mummery et al. (1999).
Unfortunately, despite being well rested prior to the session,
patient JB exhibited gross spontaneous (seemingly arbitrary)
movements during the PET scanning session. Although 12
scans were obtained (six experimental, six control) and the
data were analysed, the movement indices generated by SPM
indicated that the results of any statistical analyses should
be considered, at best, to be extremely unreliable. In any
event, subtraction of control (signal-correlated noise stimuli)
from test stimuli (speech) revealed no significant foci of
activation.
Discussion
In this exploratory study, we investigated how functional
neuroimaging might be used to investigate residual cortical
processing in patients diagnosed with PVS. Two established
sensory paradigms, familiar face perception and speech
recognition, were employed in three PVS patients who had
otherwise shown no evidence of sustained, reproducible,
purposeful or voluntary behavioural responses to visual,
auditory, tactile or noxious stimulation. In two cases (KB
and DE), rCBF responses were observed which closely
resembled those seen in healthy control subjects carrying out
identical tasks. In the third patient (JB), logistic complications
Residual cognitive function in PVS 399
Fig. 2. Surface rendered normalized positron emission tomography (PET)
data from the familiar face perception task superimposed on a standard
threedimensional
magnetic resonance template (top). The subtraction shown is faces
minus control stimuli for patient KB (right) and the control volunteer
TM (left).
In both cases, strong right hemisphere activation in the fusiform
gyrus is clearly visible. The graphs below represent individual
adjusted blood flow responses
for each scan (dots) within each condition at peak coordinates within
this region. Patient KB fell asleep during three scans (labelled
'sleep') and these data
were not used in the comparison faces versus control.
Fig. 3. Surface rendered normalized positron emission tomography (PET)
data from the speech perception task superimposed on a standard
three-dimensional
magnetic resonance template. Left (bottom) and right (top) hemispheres
are shown for patient DE (right) and average data from six control
volunteers (left)
published previously by Mummery et al. (1999). The subtraction shown
is speech minus signal-correlated noise. In both the patient and the
control subjects,
strong bilateral activation in the superior temporal gyrus (anterior
and posterior sectors) is clearly visible.
during the scan acquisition period could not be overcome
and no reliable data were obtained.
In patient KB, subtraction of the control stimuli from the
familiar face stimuli yielded two significant foci of activation,
both in the right fusiform gyrus of the occipitotemporal
region (area 37). Neuropsychological evidence from braindamaged
patients has long since suggested a role for this
region in face recognition; patients suffering from prosopag400
A. M. Owen et al.
nosia (the inability to recognize faces) are usually found to
have focal damage in this cortical area. Activation in what
has often been referred to as 'the human face area', has also
been observed in many PET studies that have required
participants to view face stimuli (e.g. Haxby et al., 1991,
1994) and has been shown to be distinct from those cortical
regions associated with location perception, as well as those
mediating the perception of colour and motion. For example,
Haxby et al. (1991) examined rCBF changes while healthy
control subjects performed face matching, dot matching or a
sensory motor control task. Face matching activated occipitotemporal
area 37, bilaterally, in a region very similar to the
one activated in patient KB in the current study. In a followup
investigation, Haxby et al. (1994) used a paradigm even
more similar to that used in the current study in healthy
control subjects; volunteers were required to match faces in
one set of scans, while in control scans 'scrambled' patterns of
equivalent visual complexity were shown. The most prominent
changes in rCBF during face perception (relative to both spatial
perception and perception of 'scrambled' visual stimuli) were
observed in regions of the mid-fusiform and anterior fusiform
gyrus (areas 19 and 37), again, very close to those regions
shown to be activated in the patient described in our study.
Although none of these studies, nor recent confirmatory
studies using fMRI (Puce et al., 1995, 1996), actually prove
that the right fusiform gyrus is involved in face processing
per se (for example, similar signal changes might well have
been observed with any class of three-dimensional object),
they do confirm that this region is more involved in the
processing of meaningful objects than in the processing of
nonsense visual patterns or letter strings. Kanwisher et al.
(1997) extended these findings by identifying a cluster of
voxels in the right middle fusiform gyrus that was consistently
more active in 80% of participants when they were viewing
photographs of faces than when they were viewing photographs
of common objects. Crucially, both sets of stimuli
used were similar along many dimensions that are probably
important for low- and mid-level vision, including that they
both depicted interpretable, meaningful three-dimensional
entities. The face object comparison was then used to
define functionally a putative 'fusiform face area' in each of
the participants and activity in this region was examined
during additional tasks that involved different kinds of stimuli.
For example, the hypothesis that the fusiform face area
might be involved in distinguishing between exemplars of a
homogeneous object set was examined by comparing
responses to faces with responses to houses. Despite the fact
that both faces and houses are single basic-level object
categories, the fusiform face area response to faces was much
greater. Similarly, the fusiform face area response to pictures
of faces was found to be significantly greater than the
response to pictures of hands, confirming that the fusiform
face area does not simply respond equally to any part of the
body [see also Kanwisher et al. (1999)]. When considered
together, these studies provide convincing evidence that the
region of the right fusiform gyrus activated in patient KB is
selectively involved in 'face perception', and concurs fully
with results derived from neuropsychological data in patients.
More importantly, perhaps, the pattern of activation
observed in KB was both qualitatively and quantitatively
similar to that observed in a single, healthy control volunteer
(TM) selected for comparison with KB and scanned during
an identical paradigm. Although anatomical and global blood
flow factors precluded direct statistical comparisons between
the two subjects, examination of Fig. 2 reveals a startling
similarity between the activation observed in the right fusiform
gyrus in both cases. Significant activation foci were
also observed in control subject TM in a similar region of
the left fusiform gyrus as well as in primary visual cortex.
While these additional changes in rCBF might indicate less
widespread recruitment of regions involved in face perception
in patient KB, one should be cautious about drawing such
strong conclusions based on qualitative comparisons between
<<<Additional text truncated.>>>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
All-in-one: antivirus, antispam, firewall for your PC and PDA. Buy Trend Micro
PC-cillin Internet Security
http://www.houseoffusion.com/banners/view.cfm?bannerid=60
Message: http://www.houseoffusion.com/lists.cfm/link=i:5:151204
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/5
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:5
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.5
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54
